Nearly half (48%) of the world suffers from a headache disorder, with 11% suffering from migraine headache. Because we do not know what causes headaches, treatment either consists of treating attacks once they occur, or attempting to reduce the frequency of attacks using daily prophylactic medications or behavioral therapies. For the vast majority of individuals, management typically involves waiting for an attack to begin and then treating the attack with medication. We have begun to show that individual headache attacks are predictable using a stress-arousal model, and that this information could be used to forecast headache risk to an individual sufferer over time. Building forecasting capabilities will require both an enhanced understanding of the causes of headache, and substantial experience in applying a forecasting algorithm across a diverse range of headache sufferers. These capabilities could be used to reduce headache burden, enhance treatment, and create more research opportunities for the study of headache. This application has the following aims: Aim 1: To formally examine the stress-arousal model in eliciting headaches using the Trier Social Stress Test (TSST) and a Fasting Test (FT). Aim 2: To examine the utility of a headache forecasting algorithm in a diverse group of headache sufferers.